How do I automate lead follow-up for my small business?

Last updated 11 September 2026. Written by Oyekola Obajuwon, founder of Growveloper, who builds the website, the marketing and the follow up for owner run businesses in Nigeria, the United States and the United Kingdom.

Automate lead follow up by wiring five things in order: the form, an instant alert to a human, an automatic acknowledgement with a booking link, one record every enquiry lands in, and a reminder if nobody replied. Start with the alert. Most missed leads are missed because nobody knew they had arrived.

That is the order. The rest of this page is what each link does, what it costs, and the part almost every guide on this subject gets backwards.

What does lead follow up automation actually mean?

It means an enquiry cannot arrive at your business without something happening to it.

Not a robot pretending to be you. Not a nine email nurture sequence. The thing being automated is the handoff: from the moment somebody fills a form or sends a message, to the moment a named human knows about it and the sender knows they have been heard.

Most articles on this subject are written for a sales team with a pipeline and a dialer. If you are an owner who answers enquiries between jobs, the useful version is smaller and it works better. You are not trying to contact more people faster. You are trying to stop losing the ones who already put their hand up.

Why do leads get missed when nobody is being lazy?

Because arriving and being noticed are two different events, and only the first one is usually wired.

An enquiry lands in an inbox. Your inbox does not ring. It fills. Somewhere between that form submission and Tuesday morning, the person who sent it messaged two other businesses, and one of them replied while your notification was still an unread line in a list of forty others.

Nobody decides to answer in three days. The gap is not a decision. It is a missing wire.

The research on this is older than most of the software sold to fix it, and it is worth reading properly rather than in the version that circulates on vendor blogs.

Two things about that table. The study is from 2011, and the exact figures deserve care, because phones, forms and patience have all changed since. What has not changed is the cause. An enquiry that arrives somewhere nobody is watching gets answered whenever somebody next looks, and that is a property of wiring, not of the year.

The second thing is what is missing from the table. Search this topic and you will meet two figures everywhere: that 78% of customers buy from whoever responds first, and that replying in five minutes makes conversion a hundred times likelier. Neither traces back to a study with a stated sample. Neither appears on this page. A number that cannot be followed home has no business inside an argument that depends on being believed.

Which leaves the finding that matters, and it is not the one the software industry sells. You do not need to reply in five minutes. You need to know within five minutes that there is something to reply to. Those are different problems, and only one of them needs building. That argument, in full, is the 42 hour gap.

What should I automate first?

The alert. Before the sequence, before the CRM, before anything clever.

One rule, written as a sentence you could hand to anybody: no enquiry, from any channel, may arrive without a named person's phone making a noise within sixty seconds.

That is it. That is the whole first automation, and it is usually an afternoon of work. It does not need artificial intelligence, a subscription, or a new platform. It needs your form to push a notification to a phone rather than only to an inbox.

Do this first because it is the only step that changes this month's numbers rather than next quarter's, and because every later step assumes it. A booking link is worthless if nobody notices the booking. A CRM is a filing cabinet for leads you already lost. For a trade where three firms get called about the same job, being the business that answers first is most of the competitive advantage available, and it is bought with a notification rather than a bigger advertising budget.

What are the five links between an enquiry and a job?

Think of it as a chain rather than a funnel, because a chain fails at exactly one place and you can go and look at which.

The reason to build them in that order is that each one makes the next one measurable. Wire the alert and you learn your real response time. Wire the acknowledgement and you learn how many people book without being chased. Wire the record and you learn which channel is worth money. Skip to the record first and you have built a very tidy account of leads you are still losing.

The glue between the links is the boring part and it is where most of the value sits. Zapier will connect almost anything to almost anything and is the right answer until volume, branching or its own pricing starts to bite. Make handles more complicated branching for less money and asks more of you. n8n is the one to reach for when the workflow has to live on your own infrastructure or touch data you would rather not hand to a third party. All three do the same job here. None of them is a strategy.

What should the automatic reply actually say?

Short, human, and specific enough that it could not have been sent to anybody else.

The job of this message is not to sell. It is to stop the person shopping. It has to do three things: confirm you received the specific thing they asked about, tell them when a human will answer, and give them a way to take the next step without waiting for you.

Hi Sarah, thanks for getting in touch about the three bedroom repaint in Lekki. I have your message and I will come back to you personally with a quote before the end of tomorrow. If it is easier to talk it through, you can grab a time that suits you here, and if you have photos of the rooms, replying to this message with them will speed the quote up.

Read it out loud. Nothing in it claims to be automated and nothing in it pretends not to be. It names the job, it commits to a deadline you can actually keep, it offers a booking link, and it asks for the one thing that makes your real reply faster.

Two rules that save this from being spam. Promise a time you will hit even on a bad week, because a broken promise is worse than a slow reply. And never send more than one automatic message before a human has said something, which is the line between an acknowledgement and a sequence nobody asked for.

Do I need a CRM for this?

No, and buying one first is the most common way this project stalls.

What you need is one place every enquiry lands, with its source attached, and one person accountable for each one. That can be customer relationship management software. It can equally be a shared inbox with rules, or a spreadsheet that gets written to automatically. The tool matters far less than the fact that something writes to it without a human retyping anything.

The test is not which product you chose. It is whether you can answer, without opening four apps, how many enquiries came in last month, where they came from, and how many are still waiting. If you can answer that from a spreadsheet, the spreadsheet is a CRM. If you cannot answer it from your CRM, it is a filing cabinet.

Buy the software when the manual version is working and its limits are the thing slowing you down. That is the same reasoning as ranking automation ideas by boredom: automate the job you already do badly and often, not the one that sounds most advanced.

What about leads that come from WhatsApp, Instagram or the phone?

Then those are your front doors, and the failure is identical. A message inside one app is seen only when somebody opens that app.

RideOn Nigeria is a premium chauffeur and car rental service in Lagos, and when the work started its bookings arrived over WhatsApp. Drivers onboarded through a basic form on a static site. The partners supplying vehicles had no portal at all. Admin lived in spreadsheets and chat threads, and the owner was running the whole thing from Canada.

The tempting move is to build an app and make everybody move to it. That is a development project, and it usually ends with a channel nobody uses and a WhatsApp thread that carries on regardless. The decision that worked was the other one: keep the door people were already walking through, and build the business behind it. Four user types, four portals, mapped end to end before any code, web first rather than native so there were no app store cycles to wait on. When it launched, drivers and customers used it without a marketing push, because it was built around what they were already doing. The whole account is in the RideOn case study.

The lesson for a business with four front doors is not to close three of them. It is to make all four ring the same bell. A WhatsApp Business account can trigger the same alert and write to the same record as your website form. So can a missed call. Route every channel to one destination, then raise the alarm on the destination rather than on each app.

What does this cost, and how long does it take?

The honest answer has two halves, and the first one is free.

The alert, the acknowledgement and a booking link can be built with tools most businesses already pay for, in an afternoon, by anybody comfortable clicking through a settings page. There is no reason to wait for a budget for that, and if somebody quotes you for it as a project, they are charging you for an afternoon.

The second half is when the chain has to be real: several channels, a record that stays accurate, quoting rules, and a handoff that survives you being on site all day. Growveloper quotes those per request rather than by the hour, and they usually go live in one to four weeks each, with the scope and the process published on the AI and automation page. Retainers for the ongoing version start at $1,500 a month. If the honest first question is which of these you actually need, that is a free consultation rather than a purchase.

The comparison worth running is not software price against agency price. It is what one lost job is worth to you, multiplied by the number of enquiries last month that nobody answered. Most owners have never calculated that number, and it is almost always bigger than the build.

What should stay human?

More than the software industry would like you to believe.

The quote stays human, or at least the number does. A form can collect everything needed to price a job and a workflow can draft it, but somebody who understands the work should look before it goes out, because the cost of an automatically wrong price is a job done at a loss or a customer who stops trusting you.

The judgement call stays human. Whether to take an awkward job, whether the timeline is real, whether this client is going to be difficult. No workflow knows any of that.

And the actual conversation stays human. What is being automated is the gap before the conversation, not the conversation. An owner who replies personally within the hour, every time, with a system making sure they know to, beats any sequence ever written for a business this size. The automation exists so that your attention lands where it is worth something, which is the same argument as the integration layer made about software generally.

How do I know it is working?

There is a test that takes ten minutes and costs nothing, and it is more useful than any dashboard you could buy.

Stop reading. Fill in your own contact form. Then send your business an Instagram message. Then a WhatsApp message. Then leave yourself a voicemail. Do it on a normal working day, at a moment you are not sitting there waiting for it.

Now write down, for each one, how long it took to reach you. Not how long it took you to reply. How long it took to find out. That number is your real response time, and it is the only one worth having.

Then run it again in a fortnight, once the alert is wired, and watch the biggest number collapse. That is the whole proof. Most owners are surprised by at least one channel, and it is usually the one they are paying to advertise.